What LJG Card does
LJG Card is a specialized tool designed for converting textual content into visually appealing PNG images using generated raster graphics and HTML typography. It supports four distinct templates: long reading cards, multi-card series, comic formats, and whiteboard layouts, allowing users to choose the format that best fits their content needs. The skill is triggered by specific commands in Chinese, making it accessible for users fluent in the language, while also being useful for those working with visual content in a variety of contexts.
The process begins with the user providing text input, which can be sourced from URLs, pasted directly, or read from local files. The skill then processes this input through a series of steps, including content extraction, visual theme establishment, and image generation. It utilizes image generation tools to create a representative image that aligns with the content's semantics and style before producing the final output. The generated images are saved as PNG or JPG files, ensuring high quality and adherence to specified dimensions.
Each template has specific parameters that dictate the output's structure, such as size and visual elements. For instance, the long card mode is designed for single images with structural anchors, while the comic mode focuses on a series of panels. This flexibility allows users to tailor their visual outputs to their specific communication needs, whether for educational materials, presentations, or creative projects.
LJG Card is particularly suited for designers, educators, and content creators who need to transform written information into visual formats efficiently. By automating the image generation process, it saves time and enhances the visual appeal of presentations and educational content, making complex ideas more accessible and engaging.
When to use it
Use LJG Card when you need to convert text into visual formats for presentations, educational materials, or creative projects.
When not to use it
This skill may not be suitable for users looking for advanced graphic design capabilities or those who require extensive customization beyond the provided templates.
What you can build with it
Creating Educational Materials
Use LJG Card to convert lecture notes into visually engaging long cards for student handouts.
Designing Marketing Content
Generate multi-card series for social media campaigns to showcase product features in a visually appealing format.
Developing Comic Strips
Transform narrative text into comic panels, making storytelling more engaging and accessible.
How to install LJG Card
View source1. Install with the skills CLI
npx skills add lijigang/ljg-skills/ljg-card --agent claude-code2. Or install it manually
Download the skill folder and drop it into ~/.claude/skills/ for all projects, or .claude/skills/ to scope it to one repo. Restart Claude Code so it picks up the new skill.
Anthropic's agentic coding CLI, and the reference implementation of Agent Skills. Drop a skill folder into ~/.claude/skills and Claude Code loads it automatically whenever a task matches the skill's description. Claude Code docs
Inside SKILL.md
Written by lijigangljg-card:铸
内容进去,PNG 出来。生成图负责把思想变成可见动作,HTML 负责把话说准确;模具只决定两者怎样相遇。
参数
| 参数 | 模具 | 尺寸 | 图像角色 |
|---|---|---|---|
-l(默认) | 长图 | 1080 × auto | 1–3 个结构性视觉锚点 |
-m | 多卡 | 1080 × 1440 | 同系列母题;每卡最多一幅主图 |
-c | 漫画 | 1080 × auto | 所有分格的漫画主画面 |
-w | 白板 | 1080 × auto | 概念隐喻与局部手绘物件 |
未给参数时使用 -l。
必读顺序
每次执行都必须依次 Read:
references/taste.mdreferences/image-generation.md- 当前 mode 文件
- 当前 HTML 模板
不得跳过共享图像协议直接写提示词,也不得把一种 mode 的图像语法套给另一种。
共同生产线
- 读取 URL、粘贴文本或本地文件,确认标题、作者、来源与事实边界。
- 提炼内容判断,建立视觉母题表:判断 → 冲突 → 视觉动词 → 承载物 → 安全区。
- 调用当前环境的 image generation 工具,先生成一张代表图校准语义与系列风格;通过后才扩展其余图片。
- 将图片保存为本地 PNG/JPG,逐一核对文件、尺寸、构图、无字与来源属性。
- 将所有可读文字、数字、公式、标签、箭头和来源放入 HTML/CSS;图片只承担场景与隐喻。
- 读取对应模板,替换全部占位符。无图槽必须显式设为
data-state="empty";有图槽必须提供本地路径和语义化alt。 - 截图前等待字体与全部图片加载成功;任一图片损坏就停止。
- 交付前检查整图;长图再按顶部、中段、底部做重叠分段检查。
关键生成图失败时最多做两次定向重生。仍失败就说明阻断原因,不得改用远程占位图、伪图标或矢量图悄悄兜底。
输入与命名
- URL:用当前可用的网页读取工具获取正文,并保存明确来源。
- 粘贴文本:直接使用,不补写原文没有的事实。
- 文件路径:Read 本地文件。
{name}:从标题或核心判断提取,中文可保留,去标点,最多 20 个字符。
截图工具
从 skill 根目录运行:
bun assets/capture.ts <html> <png> <width> <height> [fullpage]
依赖缺失时:
bun install
bunx playwright install chromium
截图脚本会等待字体与本地图片。不要绕过它的加载门禁。
Footer
-l、-c、-w:左侧保留 logo + 李继刚;右侧用{{SOURCE_LINE}}写明确来源,没有来源则替换为空字符串。-m:保留页码信息,不新增来源推断。- logo 是既有品牌位图,不属于生成图,也不能拿来充当测试外的内容插图。
mode 路由
| 参数 | mode 文件 | 模板 |
|---|---|---|
-l | references/mode-long.md | assets/long_template.html |
-m | references/mode-poster.md | assets/poster_template.html |
-c | references/mode-comic.md | assets/comic_template.html |
-w | references/mode-whiteboard.md | assets/whiteboard_template.html |
交付合同
最终回复至少报告:PNG 绝对路径、像素尺寸、内容来源、使用的 mode、生成图数量,以及整图/分段视觉 QA 结果。若输入来自已验收 Org,先记录其路径与 SHA-256,制卡后再确认源文件哈希未变。
维护自检
升级模板或 mode 后运行:
bun run audit
bun run fixtures
第一条检查共享协议、四路引用、位图槽、空槽与禁用项;第二条在 /tmp/ljg-card-v6-fixtures/ 生成四份最小代表 HTML,随后用 capture.ts 实际截图并读回 PNG。
Frequently asked questions about LJG Card
Similar skills
Excalidraw Diagram Generator
Transform natural language into Excalidraw diagrams.
Draw.io Diagram Generator
Easily create and validate draw.io diagrams.
Image Annotations
Easily annotate images with callouts and highlights.
Banner Design
Create stunning banners for any platform effortlessly.
Scientific Schematics
Generate publication-quality scientific diagrams effortlessly.
Infographics
Create professional infographics with ease and accuracy.

